Skip to the content.

Liu Zifan

Education Experience

2016.09-2019.06 Hebei Dingzhou Senior High school

National College Entrance Examination: score: 662, rank top 0.2% in Hebei province

2019.09-2023.06 The Chinese University of Hong Kong, Shenzhen

Major: Electronic Information Engineering Degree: Bachelor

• GPA&Rank:3.767/4.0(EIE Rank 6/101)

• Scholarship:Admission Scholarship: 60,000 CNY, Academic Scholarship: 20,000 CNY, College Scholarship: 1000 CNY

• Core Curriculum:Operation System, Data Structure, Computer Program Paradigm, Computer Architecture, Distributed System and Parallel Computing, Fundamentals of Artificial Intelligence, Database System Concepts

Project Experience

Energy Internet Lab of CUHKsz                               2021.05-now

Research assistant

Project name: Application of knowledge assistance deep reinforcement learning in mobile energy storage system

• Constructed the model solving framework - mixed integer programming and the model free solving framework - deep reinforcement learning.

• Proposed two knowledge assistance methods to accelerate the convergence process of reinforcement learning.

• Analyzed the effects of different super parameters and different knowledge assistance methods on the results and convergence speed of reinforcement learning.

Project name: User customized safe deep reinforcement learning in mobile energy storage system

• This project is proceeding in an orderly manner

Ernst & Young global data science challenge                        2021.05-2021.07

Group leader

• Guided to the discussion direction of the team members and determined the fire prediction as image feature extraction.

• Constructed a full convolution neural network for image feature extraction.

• Proposed a new idea of clustering + polygon fitting method for feature extraction of fire predicting and compared with the result of the full convolution network.

• Won the first place in the Hong Kong division of Ernst & Young data science challenge and the top 6 in the global finals

ASC Student Supercomputer Challenge:Language challenges and cluster calibration  2020.11-2021.01

Group member

• Added word labels for cloze questions’ options

• Trained the model of Bert + linear layer fine tuning to complete the cloze task, and the final accuracy reached 93.5%

• Tested the comprehensive memory bandwidth of the cluster to ensure the efficiency of the cluster

• Won the second prize in the ASC competition of 2020-2021

National College Students Mathematical Modeling Competition(2021)            2021.09-2021.09

Group leader
• Proposed a way to use CART to study the effect of catalyst and reactant ratio on reactant conversion rate and C4 olefin yield rate

• Constructed the linear regression, quadratic fitting and logistic regression models between temperature and reactant conversion rate and C4 olefin yield rate

• Won the winning prize of Guangdong division of national college students mathematical modeling competition

Internship Experience

Corporation: SenseTime        Department: Computer Vision

Group: Reinforcement Learning                                       Time: 2021.01-2021.05

Intern of Chinese chess algorithm                                      Location: Beijing, China

• Constructed a parallel self-play model to speed up the training speed.

• Trained the model with offline method to ensure the convergence of the model.

• Compared the strength of the model under different super parameters calibrated the chess ability of the model.

Corporation: JOYY-BIGO        Department: Operation Management

Group: Hello school                                             Time: 2021.05-2021.08

Operation team leader                                        Location: Guangzhou, China

• Led a team of 20 people to manage the daily operation of Hello campus in 8 colleges and universities in Guangdong (CHUK(SZ), SUSTech, SZTU, GDOU, ZHCOJLU, Shaoguan college, DGUT and Guangzhou southern college). The number of users in each college exceeded 2000 and the daily active users exceeded 200.

• Expand and train internship members, and the team has grown from 5 to 20.

• Lead the team to complete the company’s tasks efficiently and ensure the smooth progress of the company’s activities in various colleges and universities.

Award Certificate

• The first place in the Hong Kong division of Ernst & Young data science challenge 2021 and the top 6 in the global finals

• The second prize in the ASC competition of 2021

• The winning prize of Guangdong division of national college students mathematical modeling competition 2021

• Dean list of SSE in the Chinese University of Hong Kong (Shenzhen) in 2021

• Dean list of SSE in the Chinese University of Hong Kong (Shenzhen) in 2022

• Dean list of SHAW college in the Chinese University of Hong Kong (Shenzhen) in 2020

Skill and Hobby

• Programming:Matlab; Python: MIPs, pytorch; C++: MPI, Pthread, Openmp; CUDA

• English level:Ites: 6.5; CET-4; University full English Teaching

• Hobby:Rubik’s cube, video clips, breaking dance